INDEX

Explanations

phrases related to conflicts, battles, and power struggles

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rica

-0.69

ohn

-0.69

vale

-0.65

ocument

-0.65

 recipient

-0.60

ı

-0.58

çīĪ

-0.57

panic

-0.57

 Warning

-0.56

POSITIVE LOGITS

 scraps

1.03

 supremacy

0.87

 survival

0.79

enegger

0.79

 dominance

0.78

 turf

0.78

 whether

0.77

 vain

0.76

 seams

0.76

 attrition

0.72

Activations Density 1.953%

phrases related to conflicts, battles, and power struggles

No Comments

No Known Activations